nmt-craft-value-proposition

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In Deep mode, the runtime explicitly fetches outsider-authored free text (e.g., review quotes and sources) via the Wave 0 “reviews-mining” step: “fetch reviews from … Reddit / Product Hunt / Trustpilot / Capterra” and returns “5–10 quotable quotes per competitor with source URLs,” which is then consumed by later stages (S3/S4) for hypothesis generation and competitiveness evidence.